Skip to Main Content

AI and Academic Research

A guide exploring the intersection between AI and academic research.

Definition of Prompt Engineering

The process of getting an AI tool to produce a user's desired output by choosing the most appropriate words, phrases, and symbols and inputting them in the format the tool recognizes best. These questions or instructions are referred to as prompts. Most generative AI tools provide guidance on creating prompts and it's a good idea to review this information before interacting with generative AI.

Some basic tips:

  • You can ask text and image generators to create output based on existing styles/genres, or even the work of individuals
  • Each output is unique - if you enter in the same prompt on separate occasions, you will not get the same result
  • Just as you may have to try multiple search words or phrases when researching a topic, you likely will have to refine your prompts as you interact with the AI tool - feel free to add more details, be more specific, etc., until you're satisfied with the results

Prompting Frameworks

The CLEAR Framework

When conceptualizing your prompt, it may be helpful to use the CLEAR framework. This framework was developed by Dr. Leo Lo of the University of New Mexico to increase the effectiveness and relevance of AI-generated content using a standardized model. You can access Dr. Lo's article here

  • Concise: brevity and clarity in prompts
    • Ensure your prompt does not include unnecessary components and that it is specific and geared towards an exact response
  • Logical: structured and coherent prompts
    • If there are multiple components to your prompt, make sure the relationship and progression between the distinct parts is clear
  • Explicit: clear output specifications
    • Give precise directions regarding the output format, content, and/or scope that you wish the AI model to produce
  • Adaptive: flexibility and customization in prompts
    • Be prepared to re-work your prompt based on previous results generated by the AI model
  • Reflective: continuous evaluation and improvement of prompts
    • Evaluate your own prompts as well as the output you received to determine how better to interact with generative AI

Other Prompting Resources